An Attentional-based Capsule Neural Network for Sentiment Analysis on Online Reviews

As a fundamental task of the natural language processing, sentiment analysis has drawn much attention from researchers. In-depth research has been conducted by previous work in this field and great progress has been made. However, these works fail to capture the emotional information in the text accurately, so it is still difficult to achieve satisfactory results. To address this problem, attentional capsule network (ACN) is proposed, which leverages the attention mechanism and the properties of capsule network to capture sentiment information from online reviews accurately. The experimental results demonstrate that our model is superior to other baselines.